JP6940858B2 - Paper piece handling device - Google Patents

Description

The present invention relates to a paper piece handling device into which paper pieces such as banknotes are inserted and a control method thereof.

In recent years, a paper piece handling device in which a user puts a piece of paper into the device has been used in various situations. Examples of the paper piece handling device include an ATM (Automatic Teller Machine), an automated teller machine, a cash teller machine, a printer, and a multi-copy machine.

Unlike the operation mode in which ATMs and multi-copy machines these days can be installed in traditional bank stores and specialty stores so that managers can immediately deal with problems, stores with little expertise or stores with no managers at all. The ratio of installation is increasing. Many such ATMs are often installed in one store, and once they break down, there is no alternative device. A user who visits a store without an alternative device cannot use the device and cannot perform desired processing. In addition, many of the recent installation conditions such as convenience store ATMs have problems that it is difficult to secure the time and place for maintenance work in the event of a breakdown, such as 24-hour operation, no dedicated maintenance area, and heavy traffic.

Patent Document 1 is mentioned as a related technique. Patent Document 1 describes a money deposit / withdrawal device that detects the occurrence of money jam generated inside the device by photographing the inside of the device and performing image analysis. This money deposit / withdrawal device is provided with an imaging means for at least photographing a transport path through which bills and coins pass, and stores an image at the time of jam occurrence.

Japanese Unexamined Patent Publication No. 2013-117759

A paper piece handling device in which a user throws a piece of paper into the device handles a piece of paper brought in or an uncertain piece of paper, and the user himself / herself executes the act of putting the piece of paper into the device. It has a feature that the uncertainty of is very large.

For example, in the case of ATMs, since the user "inserts" cash, foreign matter is mixed in, the setting method is defective, and many obstacles, failures, and errors occur, such as the insertion of deteriorated cash and the insertion of foreign currency. There is a possibility that it will become. The components of an ATM device may also be complex, and some of the factors illustrated are high in the failure rate of an actual ATM. It is desirable to establish countermeasures for these obstacles and reliable countermeasures.

In addition, each mechanical component such as an ATM has many variable factors related to performance such as environment, aging deterioration, and dirt, in addition to variations in component shape and composition, and there are also differences in the behavior of each device. Moreover, even if the same device is used, the behavior may differ depending on the time of day and deterioration over time.

In addition, in order to reduce obstacles by adding the above-mentioned uncertainties, it is necessary to comprehensively consider all the combinations of the above-mentioned variable factors and to evaluate and verify a huge number of combinations. However, in many cases, very expensive parts are used, combination verification considering all variations is practically impossible, and it is difficult to make the failure rate extremely close to zero by this method. it is conceivable that.

There are also issues from another aspect. There is also a problem of high uncertainty in the verification method when the current paper piece handling device fails. For example, in the case of ATMs, the state of cash media can be grasped by fragmentary history and errors based on ON / OFF information of optical sensors, handling data showing the state at the time of failure, and ambiguous hearing based on the user's memory. The cause of the failure is assumed and dealt with by using the contents. However, it is difficult to clearly determine whether the cause of the failure is the effect of setting the medium, the effect of the deteriorated medium, the effect of the device alone, the effect of the installation environment, or the combination of each. Similarly, there is also a problem that reproduction verification tends to be uncertain and inefficient, such as verification in an environment different from the actual situation or an evaluation machine based on uncertain information.

Patent Document 1 mainly discloses a deposit / withdrawal device such as POS (Point Of Sales) for depositing / withdrawing money, banknotes, and the like. In Patent Document 1, a plurality of cameras are installed on the transfer path of money and banknotes from the deposit / withdrawal port to the storage unit, and the occurrence of jam of money and banknotes is monitored by these cameras.

However, Cited Document 1 does not describe at all that jam occurs depending on the state of the user's banknote set when the banknote is inserted and the state of the banknote itself such as the folding habit. That is, Cited Document 1 does not describe a specific problem that occurs when a user throws in a piece of paper.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a paper piece handling device capable of verifying a problem of handling a piece of paper and suppressing a problem of handling a piece of paper at the time of inserting the piece of paper.

The paper piece handling device according to the embodiment of the present invention includes a paper piece loading port and a transport path, and at least photographs the transport mechanism unit for transporting the paper pieces loaded from the loading port by the user and the loading port, and the above-mentioned From the first camera unit that acquires an image showing the state of the paper piece caused by the user input from the insertion slot, and the image group taken by the first camera unit each time the paper piece is handled, the said A learning unit that extracts user-side factor information caused by the user input from the slot and machine-learns the state of the piece of paper as learned data from the user-side factor information, and a learning unit that generates the information. When the learned data is compared with the user-side factor information at the time of user input and the characteristics match, the difference information between the state of the paper piece inserted from the input port and the standard state is associated with the corresponding image. It is provided with a control unit that accumulates in the storage means.

The control method of the paper piece handling device according to the embodiment of the present invention is the use of the image group collected by taking a picture every time the user handles the paper piece at least by taking a picture of the slot for inserting the paper piece. The user-side factor information caused by the person is extracted, and the state of the paper piece is machine-learned in advance as learned data from the user-side factor information. The process of detecting feature matching is executed by comparing the data with the state of the piece of paper that has passed through the transport path this time, and the learned data is compared with the user-side factor information from the time of inputting the user to match the features. In this case, the difference information between the state of the piece of paper loaded from the slot and the standard state is associated with the corresponding image and stored in the storage means.

The program for a paper piece handling device according to an embodiment of the present invention includes a paper piece loading port and a transport path, and at least photographs a transport mechanism unit for transporting a paper piece loaded from the loading port by a user and the loading port. Each time a piece of paper is handled, the control computer of the piece of paper handling device including the first camera unit for acquiring an image showing the state of the piece of paper caused by the user thrown in from the insertion port is used. From the image group taken by the camera unit of the above, the user-side factor information caused by the user input from the input port is extracted, and the state of the paper piece is used as learned data from the user-side factor information. When the learning unit for machine learning, the learned data generated by the learning unit, and the user-side factor information from the time of inputting the user are compared and the characteristics match, the piece of paper inserted from the input port is used. It is characterized in that it operates as a control unit that associates the difference information between the state and the standard state with the corresponding image and stores it in the storage means.

According to the present invention, it is possible to provide a paper piece handling device that realizes verification and / or suppression of defects in paper piece handling through machine learning.

It is a functional block diagram which shows the paper piece handling apparatus 1 of 1st Embodiment. It is a flowchart which shows the processing operation example of the paper piece handling apparatus 1 of 1st Embodiment. It is a flowchart which shows the processing example of the machine learning for the paper piece handling apparatus 1 of 1st Embodiment. It is a flowchart which shows the processing operation example of the paper piece handling apparatus 1 of 1st Embodiment. It is a flowchart which shows the processing example of the machine learning for the paper piece handling apparatus 1 of 1st Embodiment. It is a functional block diagram which illustrates the maintenance system including the ATM 100 of the 2nd Embodiment. It is explanatory drawing which shows typically the mode that the transport mechanism part 10 of the cash processing apparatus 2 of 2nd Embodiment collects information about the state of the banknote in each unit as image data. It is explanatory drawing which shows typically the operation which the transport mechanism part 10 of the cash processing apparatus 2 of 2nd Embodiment compensates transport according to the state of the banknote in each unit. It is explanatory drawing which showed various information and processing which are related to the iterative learning specialized for the banknote of the cash processing apparatus 2 of 2nd Embodiment.

Embodiments of the present invention will be described with reference to the drawings.

[First Embodiment]
FIG. 1 is a functional block diagram showing a paper piece handling device 1 according to an embodiment of the present invention. The paper piece handling device 1 includes a transport mechanism unit 10, a camera unit 20, a learning unit 30, and a control unit 40.

The transport mechanism unit 10 includes a paper piece loading port and a transport path, and is configured to transport the paper piece inserted into the paper piece loading port to a predetermined position in the apparatus. For example, the transport mechanism unit 10 includes a loading port for loading paper pieces by the user, a paper piece stacking unit for accumulating the loaded paper pieces, and a transport path between them, and the transport path is composed of several mechanical mechanisms. be able to. Further, the transport mechanism unit 10 may be configured to return the loaded paper piece to the user as it is, or to process and return it as needed. Further, the transport mechanism portion 10 contains a movable portion that is in charge of transporting a piece of paper. This movable unit operates when a piece of paper is conveyed to a predetermined position in response to control from the control unit 40 or the like by a motor or the like.

Further, the transport mechanism unit 10 may include a device control unit that controls the movable unit. This device control unit operates a movable unit when a piece of paper is conveyed to a predetermined position in response to control from the control unit 40 or the like. Further, the device control unit controls each movable part based on a compensation value for finely adjusting the speed, rotation speed, movable length, angle, etc. of each movable part when transporting the loaded paper piece. It may be configured.

The camera unit 20 (20-1, ... 20-n) photographs the transport path in the device and the slot into which the user inserts a piece of paper, and records the image in a required storage means. The camera that captures this slot is the first camera unit. With this first camera unit, it is possible to take an image showing the state of a piece of paper caused by the user inserted into the insertion slot. Further, the camera that captures the transport path becomes the second camera unit. With this first camera unit, it is possible to take an image showing the state of a piece of paper caused by the device. The shooting may be a moving image, a still image, or a composite thereof. In the above example, the camera unit 20 may photograph the input port, the transport path, and the paper piece accumulating unit, respectively, and record each image in the required storage means. This storage means may be appropriately used in any or a combination of, for example, the semiconductor storage provided in the camera unit 20, the storage provided in the device, or the network storage. The camera unit 20 appropriately sets the shooting range so as to comprehensively cover the movement path of the piece of paper. Further, the camera unit 20 may be provided with a reflective surface with a reflective material or paint having a mirror function so as to enhance the photographing point used for machine learning in the transport mechanism unit 10. Similarly, the camera unit 20 may use a fisheye lens to enhance the visible range so as to enhance the imaging point used for machine learning in the transport mechanism unit 10.

The learning unit 30 machine-learns an image of the state of a piece of paper passing through at least the slot and the transport path, which is taken by the camera unit 20 collected every time the piece of paper is handled, together with the handling data. In machine learning, the state, condition, likelihood, etc. of a piece of paper that causes an error, the state, condition, likelihood, etc. of a piece of paper that does not cause an error are collected as learned data. In this embodiment, the user-side factor information indicating the state of the piece of paper generated by the user and the device-side factor information indicating the state of the piece of paper caused by the device itself are collected and used for learning.

Further, the learning unit 30 may reflect the state of the piece of paper in the learning result when the state of the piece of paper is obtained from the existing sensor value, flag, state signal, or the like without obtaining the state of the piece of paper only from the image. The state of the piece of paper obtained from each image may be captured as data, for example, the difference from the standard state of transport (ideal state of transporting the piece of paper, initial design state).

Further, it is desirable that the learning unit 30 learns the factor information at the time of transporting a piece of paper regarding the difference from the standard state for machine learning by dividing it into levels. This makes it easier to reflect, for example, the likelihood of the pattern of the state in which an error occurs or the state before the error occurs in the control of each part, information analysis, and information sharing.

Further, as will be described later, the learning unit 30 may be configured to learn the compensation value that brings the transport for the movable unit closer to the standard state and include it in the learned data. It is considered that the better condition of the piece of paper has more influence on the optimization of the compensation value by learning. For this reason, by incorporating the state (image) in which a piece of paper is inserted into the insertion slot into the learning information together with the handling data, it is easy to obtain learning results for suppressing the occurrence of an abnormal state that occurs after the insertion slot. Conceivable.

Using this learning result, the paper piece handling device 1 can collect information with higher verifiability when there is a problem in handling the piece of paper. Further, using this learning result, the paper piece handling device 1 derives an abnormal factor during transportation, and makes it possible to suppress the occurrence of the abnormal factor during transportation.

Further, when the paper piece accumulating portion is provided and the image is taken, the paper piece handling device 1 may perform machine learning by adding the accumulated paper piece image obtained by photographing the paper piece accumulating portion. Further, the paper piece handling device 1 may be relearned every time the paper piece is handled.

The handling data includes the values, flags, and operating conditions of various sensors that can be acquired on the device side during a series of flows including the insertion of a piece of paper into the slot, such as for each ATM transaction or printing by a printer. Various information with time series as needed. Further, it is desirable that the handling data include device information such as a unique number of the device itself and a connected unit number. In addition, it is desirable to capture environmental information such as temperature and humidity inside and outside the device as handling data.

When the control unit 40 handles the paper pieces, the control unit 40 digitizes the state of the paper pieces passing through each part of the transport mechanism unit 10 from the images taken by the camera unit 20 and stores them in the storage means. For the state of the piece of paper, for example, the difference from the standard state of transportation may be converted into data. By using this difference for learning and analysis, it is possible to reduce the amount of information and improve the analysis speed.

Further, when an error occurs when handling a piece of paper, the control unit 40 stores the handling data when handling the pieces of paper before and after the error, and the image taken by the camera unit 20 in association with the storage means. .. Further, in parallel with the above processing, the control unit 40 compares the learned data generated by the learning unit 30 with the state of the piece of paper that has passed the transport path this time, and when the characteristics match, before and after that. Each image taken by the camera unit 20 is accumulated together with the handling data when the piece of paper is handled. Further, when the control unit 40 stores the image, the control unit 40 stores the difference (difference information) between the state of the paper piece passing through the transport mechanism unit 10 and the standard state according to the storage means.

The image (still image, moving image, or combination) is used to generate the difference information obtained by extracting the difference between the state of the piece of paper being conveyed and the standard state. For the difference information, for example, a pattern as a set of feature points, feature lines, and feature planes obtained by image analysis can be identified from each image information, and the difference can be extracted for each factor by comparing with the pattern in the standard state. good. Examples of factors include the state of the piece of paper such as the shape and size of the piece of paper, the state of the user setting the piece of paper in the slot, the state of feeding out from the slot, the state of transporting the piece of paper, and the state of accumulating the pieces of paper.

Further, the control unit 40 has a configuration for learning the outer shape, bending habit, length in the thickness direction, set misalignment, foreign matter mixing, foreign matter adhesion, etc. of the paper piece inserted into the insertion port as a difference from the standard state for machine learning. You may adopt it. Furthermore, the sluggishness of the paper pieces during transportation, the distance between the paper pieces, the skew, and the accumulation state at the time of accumulation, the accumulation empty area, the occurrence of folding, the feeding time, the medium skew ratio, and the deformed shape are the important factors for the transportation of the paper pieces. The learning unit 30 may be used for learning.

Further, the control unit 40 extracts the difference between the error state and the state before the error as an important factor regarding the transportation of the paper piece regarding the difference from the standard state used for machine learning, and extracts the required difference. It may be recorded in a storage means in a timely manner and learned by the learning unit 30. Similarly, each of the extracted differences is associated with the image information as the respective difference information and recorded in the required storage means in a timely manner.

The image information and the difference information, which were the transfer of the paper pieces without any problem, may be deleted as appropriate, but may be used for machine learning in order to obtain the learning result leading to the more normal transfer of the paper pieces.

In addition, when one or more factors represent characteristic differences, or when an error occurs or a difference that leads to an error is recognized, each image related to the handling of the piece of paper, each difference, and a certain number before and after that. The data may be configured to be stored in any storage means for reflection in analysis and machine learning.

The stored images and handling data are stored so that the administrator can access them for appropriate use for maintenance and improvement. For example, when an error occurs, or characteristic difference information and its image group can be configured so that the information can be immediately and timely expanded to the related maintenance personnel and technicians via the network. good.

Next, an operation example of the control unit 40 of the paper piece handling device 1 of the present embodiment will be described.

FIG. 2 is an example of a flowchart showing an operation related to information collection of the present embodiment.

The paper piece handling device 1 (control unit 40) acquires (collects) camera images such as a paper piece insertion slot and a movement path each time the paper piece is handled (S101).

Next, the paper piece handling device 1 (control unit 40) detects an error (state of the paper piece causing a problem) by comparing the camera image handled this time with the learned data, as well as detecting an error based on the sensor value or the like (the state of the piece of paper causing a problem). S102).

Next, the paper piece handling device 1 (control unit 40) acquires and collects an image and difference information at the time of an error / abnormality together with the handling data (S103).

Next, the paper piece handling device 1 (control unit 40) provides the administrator with an image at the time of an error / abnormality (S104).

Further, FIG. 3 is a flowchart showing an example of machine learning processing related to information collection of the present embodiment.

The paper piece handling device 1 (control unit 40) sequentially collects an image group for each past piece of paper handled for each device, various sensor values, and the difference between the state of the piece of paper and the standard state (S201).

The paper piece handling device 1 (control unit 40) extracts features (parameters) from each of the collected data and sequentially stores them (S202).

The paper piece handling device 1 (learning unit 30) machine-learns the relationship of the feature (parameter) group that leads to the problem at the time of transporting the paper piece for each factor (S203).

The paper piece handling device 1 (learning unit 30) accumulates machine learning results as learned data for each factor (S204).

By operating as shown in FIGS. 2 and 3, the paper piece handling device 1 of the present embodiment can collect data in which the amount of information on the trouble of handling the paper piece is condensed through machine learning.

FIG. 4 is an example of a flowchart showing an operation related to abnormality suppression of the present embodiment.

The paper piece handling device 1 (control unit 40) detects the state of the paper piece by comparing the camera image handled this time with the learned data, as well as detecting an error based on the sensor value or the like (S301).

Next, the paper piece handling device 1 (control unit 40) adjusts each parameter of the movable portion of the transport path in real time according to the state of the paper piece such as the insertion slot accompanying the handling of the paper piece (S302).

Further, FIG. 5 is a flowchart showing an example of machine learning processing related to abnormality suppression of the present embodiment.

The paper piece handling device 1 (control unit 40) sequentially collects an image group for each past piece of paper handled for each device, various sensor values, and the difference between the state of the piece of paper and the standard state (S401).

The paper piece handling device 1 (control unit 40) extracts features (parameters) from each of the collected data and sequentially stores them (S402).

The paper piece handling device 1 (learning unit 30) derives each compensation value for the transport mechanism unit 10 that brings the factors leading to the problem in transporting the paper piece closer to the standard state (S403).

The paper piece handling device 1 (learning unit 30) accumulates machine learning results as learned data for each factor (S404).

By operating as shown in FIGS. 4 and 5, the paper piece handling device 1 of the present embodiment can be operated so as to suppress the occurrence of a defect when handling the paper piece through machine learning.

Next, a more specific embodiment of the present invention will be described.

[Second Embodiment]
The present embodiment assumes an ATM 100 that handles banknotes as a piece of paper.

FIG. 6 is a functional block diagram illustrating a maintenance management system including the ATM 100 of the present embodiment.

The ATM 100 of this embodiment includes a cash processing device 2 and is connected to a communication network. The device configuration of the cash processing device 2 is the same as that of the paper piece handling device 1 of the first embodiment, and includes a transport mechanism unit 10, a camera unit 20, a learning unit 30, and a control unit 40. Further, the cash processing device 2 is provided with a storage means for storing various data including learned data and images of each part. Further, the cash processing device 2 has a camera unit that captures a transfer mechanism unit 10, an insertion port for directly inserting and removing banknotes by the user, a feeding unit for taking out banknotes one by one from the insertion port, a transfer path, and a paper piece accumulation unit. 20 is provided.

The transport mechanism unit 10 of the present embodiment includes a movable unit 11 that enables the behavior of the bill to be compensated by individual mechanisms in the bill unit, and a device control unit 12 that controls the compensation of the movable portion 11.

The device control unit 12 optimizes the feeding gap, the medium pressing force, the transport speed, the guard time, and the movable mechanism of the integration area in real time according to the learned data.

Further, the camera unit 20 is composed of a plurality of camera units (20-1 to 20-n), and is in a set state by a banknote user, a payout state from a banknote slot, a banknote transport state, and a banknote. Each image such as the accumulated state of the banknotes is captured. In this way, by arranging the camera unit 20 capable of monitoring the state of banknotes in each unit in each main part, it is possible to collect information on the state of banknotes in each main part for all transactions. ..

The control unit 40 includes a processor 41, an image data acquisition unit 42, and a pattern recognition unit 43. The processor 41 controls the entire device and also performs integrated control according to the present invention. The image data acquisition unit 42 collects the image captured by the camera unit 20 in the storage means. The pattern recognition unit 43 refers to the image (state of banknotes) taken by the camera unit 20, recognizes the difference between each factor and the standard state at the time of operation, and generates pattern information. Collect pattern information in a storage means.

In addition, the control unit 40, as a whole, recognizes the difference information from the standard state at the time of each operation as a pattern, and classifies and manages it as pattern recognition data of each element for each transaction. The difference information here includes external shape, bending habit, length in thickness direction, set deviation, foreign matter contamination, and foreign matter adhesion as factors on the user side, and sluggishness, distance between media, as factor information on the device side. Inclined, accumulated state, accumulated empty area, generated breakage, feeding time, medium skew rate, and deformed shape may be included as factors.

Further, the control unit 40 optimizes each movable unit in real time according to the state of transporting the paper piece via the device control unit with reference to the compensation value described later.

The learning unit 30 sets the bill setting state by the user in the insertion slot and the deterioration status of the bill as the user side factor information, and sets the bill feeding state, the bill transporting state, and the bill accumulation state as the device side. Used for iterative machine learning as factor information. In this machine learning, the outer shape, bending habit, thickness direction length, set deviation, foreign matter contamination, foreign matter adhesion, sluggishness, distance between media, skew, accumulation state, accumulation of paper pieces extracted as differences from the standard state. Relationships and relationships are learned using empty areas, generated folds, feeding times, medium skew ratios, and deformed shapes as extraction factors, and learning results are generated.

Further, the learning unit 30 learns the compensation value that minimizes the difference from the standard state with respect to the state of the piece of paper in which the abnormality is not detected together with the handling data (for example, the difference pattern from the standard state related to transportation) as learned data. do. This compensation value is derived according to the movable unit 11 that can be adjusted by the device control unit 12. For example, the compensation values for adjusting the feeding gap, the medium pressing force, the transport speed, the guard time, and the movable mechanism of the accumulation area are derived as learned data.

Further, the learning unit 30 repeatedly learns the difference between the information obtained by correcting the fluctuation parameter of the mechanism by the device control unit 12 and the bill pattern data before and after the correction from the standard state in the learning unit 30, and reflects this learning result in the correction value. You may let it update the learned data.

With the above configuration, the cash processing device 2 converts the state of the bill into image data by the camera unit 20 for each transaction, and the control unit 40 converts the image data into a pattern via the image data acquisition unit 42 and the pattern recognition unit 43. Data is processed into data, and image data and pattern information are stored and collected in storage means, respectively.

Further, in this embodiment, the image data and the pattern data collected in the storage means can be transmitted to the data server 110 via the network. Image data and pattern data are provided to a management department 120 including a maintenance department 121 connected to a network, a technical department 122, a data analysis department 123, and the like as needed. The management department 120 collects the same kind of information from a large number of ATMs (for example, ATMs 101 to 103) as needed, and uses it for feedback to various problems and development of a latecomer device.

Here, with reference to FIGS. 7 and 8, a state of collecting information as image data on the state of banknotes in each main unit of the transport mechanism unit 10, such as a bill slot, and an example of an operation of compensating for transport are specifically shown. explain.

FIG. 7 is an explanatory diagram schematically showing how the transport mechanism unit 10 of the cash processing device 2 collects information on the state of banknotes in each unit as image data.

Further, FIG. 8 is an explanatory diagram schematically showing an operation in which the transport mechanism unit 10 of the cash processing device 2 compensates for the transport according to the state of the banknotes in each unit.

7 and 8 show that the camera unit 20 is provided in each of the unit constituting the input port, the unit constituting the accumulation unit, and the unit constituting the transport path in the cash processing apparatus 2.

FIG. 7A illustrates the state of the banknote bundle at the time of photographing the insertion slot and starting feeding from the insertion slot. As shown in the figure, the camera unit 20-1 is arranged near the slot. The camera unit 20-1 enhances the shooting range with a reflector and a transparent guide. The reflective surface of the reflector has a mirror function. As with the mirror glass, the inner surface of the slot may be mirrored so that the inner surface of the device cannot be seen. Further, the camera unit 20-1 may be arranged so that the gap roller and the peeling roller are also photographed. FIG. 7B illustrates the state of the banknote bundle at the time of photographing the banknote collecting section and starting collecting the banknotes in the banknote collecting section. As shown in the figure, the camera unit 20-2 is arranged in the vicinity of the bill collection unit. The camera unit 20-2 uses a reflector and a transparent guide to enhance the shooting range. 7 (c) and 7 (d) show the state of banknotes at the time of photographing and transporting the transport route. As shown in the figure, the camera unit 20-n is arranged in the transport path. The camera unit 20-n enhances the photographing range with a reflector and a transparent guide. Although one camera unit 20 is shown in each drawing, it is not always necessary to have one camera unit 20. The unit may be imaged by a plurality of camera units 20, or a plurality of units may be imaged by a single camera unit 20.

FIG. 8A illustrates a picture of the slot and a transport element that compensates for the state of the bill. As shown in the figure, the state of the banknotes (banknote bundle) set in the insertion slot is photographed by the camera unit 20-1 arranged near the insertion slot, the state of the banknotes is monitored by the camera unit 20, and the banknotes (banknote bundle) are monitored. From the state (image pattern information such as bending habits and deformations), the transparent fluctuation guide is pressed, the position is corrected, the position of the feeding gap is corrected by the gap roller, and the rotation speed of the peeling roller is corrected via the device control unit 12. To enable. In addition, the fact that foreign matter is mixed may be identified as the state of banknotes (banknote bundles).

FIG. 8B illustrates a transport element that compensates for the state of the bill and the photograph of the stacking portion. As shown in the figure, the state of the bills (banknote bundles) set in the stacking unit is photographed by the camera unit 20-2 arranged near the stacking unit, the state of the bills is monitored by the camera unit 20, and the bills (banknote bundles) are monitored. From the state (image pattern information such as bending habits and deformations), it is possible to press the transparent variation guide, correct the position, and correct the rotation speed of the integrated roller via the device control unit 12. Further, it is sufficient to identify that foreign matter is mixed in the collecting portion as the state of the banknote (banknote bundle). 8 (c) and 8 (d) illustrate the image of the transport path and the transport element that compensates for the state of the bill. As shown in the figure, the state of banknotes (banknote bundles) passing through the transfer path is photographed by the camera unit 20-n arranged near the transfer path, the state of banknotes is monitored by the camera unit 20, and the banknotes (banknote bundles) are monitored. ) (Image pattern information such as bending habits and deformations), the transfer roller can be pressed and corrected and the rotation speed can be corrected via the device control unit 12. Further, it is sufficient to identify that foreign matter is mixed in the transport path as the state of the banknote (banknote bundle).

When a user inserts banknotes into the slot, the set state of the banknotes is often different from the case where the banknotes are uniformly separated and transported one by one, such as denomination, number of sheets, dirt, damage, breakage, and foreign matter contamination. Uncertainties are always present.

Also, on the unit side, there are many uncertain factors related to the same performance such as mechanical shape, roller diameter, material, aging, and environmental changes for each product, so each of the above uncertain factor factors is complicated. Due to the relationship, the behavior of banknotes is different each time.

First, the camera unit 20 monitors the state of the banknote (banknote bundle) set by the user in the slot, and further monitors the state of the banknote (banknote bundle) in the subsequent process, and the state of the banknote (banknote bundle) (image pattern information). ), It is possible to press the pressure fluctuation guide, correct the position, correct the position of the feeding gap by the gap fluctuation roller, and adjust the speed of the peeling portion roller by the speed fluctuation roller via the device control unit 12. Therefore, for each ATM device, it is possible to make adjustments related to the optimum transport of banknotes that matches the set state of the own device.

The determination of the compensation value of the device control unit 12 is as described above, and the input image information at the start of the transfer operation of the input port, and the images of the subsequent transfer process and the integration process, collected in advance by the learning unit 30. Machine learning is performed on the relationship of information in advance, and based on this learning result, compensation that optimizes the payout of the bills input this time is immediately executed. Further, the bill collecting section and the bill transport path shown in FIGS. 7 (b) to 7 (d) are also optimally compensated via the movable section 11, the device control section 12, and the learning section 30 in the same manner as the slot. By repeatedly operating machine learning, the more optimal compensation value of the movable portion 11 is learned each time. This makes it possible to achieve constant improvement while autonomously following changes over time.

Further, in the present embodiment, if a problem occurs, detailed image data at the time of occurrence and its processed data can be downloaded by the data server 110 via the network, and information can be shared in real time with related parties outside the store. become. For this reason, it is possible to facilitate quick maintenance with few mistakes in replacement parts, countermeasure tools, investigation of the cause of defects, and provisional measures.

As shown in the figure, in order to collect the state of the banknote as detailed image information, the members located between the banknote and the camera unit 20, such as the slot, the transparent fluctuation guide, the transparent guide, and the banknote collection unit, are It has a structure with enhanced visibility by using a transparent color, adding a hole shape, and providing many gaps. Further, in order to collect the state of banknotes in a wider range, it is desirable to arrange reflectors as illustrated in FIGS. 7 and 8 in various places so that the state of banknotes can be more identifiable. Further, a fisheye lens may be used in the camera unit 20 to extend the identification range of the state of the bill.

Next, various information and processing related to iterative learning specialized for banknotes will be described.

FIG. 9 is an explanatory diagram showing various information related to iterative learning specialized for banknotes and processing in the cash processing apparatus 2.

The cash processing device 2 includes handling data such as the handling date and time, the machine number, image information (banknote set state) in which the user sets the bill in the slot, which is user-side factor information (cause information), and bills on the route. The image information (deterioration state of banknotes) for each factor is used as at least input information, and the device-side factor information such as the banknote delivery state, banknote transport state, and banknote accumulation state is added, and the difference from the standard state of each factor is obtained. Collect as output information and carry out repetitive learning.

The ATM 100 (cash processing device 2) of the present embodiment has two elements of a banknote set state and a user-side factor information which is a banknote deterioration state, which could not be grasped in detail as an ATM, and after the banknotes are taken into the unit. The five elements of the three elements of the device-side factor information, which is the feeding state, the transporting state, and the integrated state, are collected as image information (still image, moving image, or combination) by the camera unit 20 arranged in each unit. Then, the cash processing device 2 uses the learned data to extract important factors and to improve by maintenance before a serious defect occurs (defect suppression step 1).

Further, the cash processing device 2 uses the iteratively learned data to control the parameters of important factors and realize autonomous improvement (defect suppression step 2). Specifically, the payout gap, medium pressing force, transport speed, guard time, and movable mechanism of the accumulation area, which affect the transport performance of banknotes, can be variably controlled as controllable factors to correspond to the corresponding pattern information. Optimized control. The above-mentioned applicable pattern information and controllable factors that affect the transport performance of bills are managed by dividing the pattern of the state before the error into levels in addition to the transaction that causes an error, so that the pattern level becomes more ideal. It repeatedly learns input information and output information, and autonomously changes it to an ideal state that operates in the same way as the standard state.

The controllable factors are the feeding gap, the medium pressing force, the transport speed, the guard time, and the movable mechanism of the accumulation area, but the controllable factors are not limited to these.

As described above, the two elements of the banknote set state of the cash processing device 2 and the user-side factor information which is the deterioration state of the banknote, and the payout state, the transport state, and the accumulation state after the banknote is taken into the unit. Each operation related to the five elements of the three elements, which is the factor information on the device side, is processed as the difference pattern information from the standard state, and the difference pattern information can be used as the input information to collect the image information of the state of the paper piece. Further, each transport mechanism can be optimized and controlled by using the difference pattern information as input information. As a result, it is possible to realize transport control that approaches the standard state under any conditions by adding many fluctuation factors by machine learning.

Furthermore, along with collecting image information in a state connected to an abnormality and optimizing control of the transport mechanism, a compensation value that minimizes the difference from the standard state of the image group using an image group in which no abnormality actually occurs is set. Generated by iterative learning. By compensating the transport mechanism portion 10 (movable portion 11) with this compensation value, it is possible to autonomously and in real time optimize the transport for each device. This optimization of transportation can always respond to variations in the user's bill setting method, variations from device to device, temperature / humidity changes, aging deterioration, and regional characteristics.

The learned data of each device is collected and managed by the management department. This makes it easier to understand problems around the transport mechanism of the device regardless of errors, and makes it easier to carry out design changes for overall optimization in a timely manner.

The iterative learning function incorporates an AI (Artificial Intelligence) function represented by deep learning, and pre-learns the compensation values of various mechanisms for the input information. This makes it possible to autonomously guide unexpected optimizations that were not intended as control targets.

As described above, the paper piece handling device to which the present invention is applied can realize verification and / or suppression of defects in paper piece handling through machine learning.

The control computer of the paper piece handling device may be realized by using a combination of hardware and software of the computer system. The computer system includes one or more processors and memory tailored to the desired form. Further, in the form of this computer system, in each part, a control program of a paper piece handling device is expanded in the above memory, and based on this program, hardware such as one or a plurality of processors is operated by an execution instruction group or a code group. By doing so, it should be realized. At this time, if necessary, this program may realize each part in cooperation with the functions provided by the software such as the operating system, the microprogram, and the driver.

The program data expanded in the memory appropriately includes an execution instruction group, a code group, a table file, content data, and the like that operate the processor as one or a plurality of the above-mentioned parts.

In addition, this computer system does not necessarily have to be constructed as a single device, but is constructed by combining multiple servers / computers / virtual machines, so-called thin clients, distributed computing, and cloud computing. May be good. Further, a part / all parts of the computer system may be replaced with hardware or firmware (for example, one or more LSIs: Large-Scale Integration, FPGA: Field Programmable Gate Array, a combination of electronic elements). Similarly, only a part of each part may be replaced with hardware or firmware.

In addition, this program may be temporarily recorded on a recording medium and distributed. The program recorded on the recording medium is read into the memory via wire, wireless, or the recording medium itself, and operates the processor or the like.

In the present specification, the recording medium includes a storage medium, a memory device, a storage device, and the like of similar terms. Examples of this recording medium include optical disks, magnetic disks, semiconductor memory devices, hard disk devices, tape media, and the like. Further, it is desirable that the recording medium is non-volatile. Further, as the recording medium, a combination of a volatile module (for example, RAM: Random Access Memory) and a non-volatile module (for example, ROM: Read Only Memory) may be used.

The present invention has been described by exemplifying embodiments. However, the specific configuration of the present invention is not limited to the above-described embodiment, and is included in the present invention even if there is a change within a range that does not deviate from the gist of the present invention. For example, changes such as separation and merging of block configurations and replacement of procedures of the above-described embodiments are free as long as the gist of the present invention and the functions described are satisfied, and the above description does not limit the present invention.

Further, in the present embodiment, the bill unit handles bills, but it may be widely applied as a coin unit handling coins or a paper handling terminal such as a printer. When handling coins, the deformation, dirt, etc. of the coin may be extracted as the difference from the standard state.

In addition, some or all of the above embodiments may also be described as follows. The following notes do not limit the present invention in any way.

[Appendix 1]
A transport mechanism unit that has a paper piece loading section and a transport path to transport the paper pieces inserted by the user, and a transport mechanism unit.
A camera unit that captures at least the transport path in the device and the input unit in which the user inserts a piece of paper.
A learning unit that machine-learns the state of a piece of paper that causes an error from the state of the piece of paper passing through the transport mechanism unit and the handling data collected by taking a picture each time the piece of paper is handled by the camera unit as learned data.
When an error occurs when handling a piece of paper, or when the learned data generated by the learning unit is compared with the state of the piece of paper that has passed through the transport mechanism unit this time and the characteristics match, before and after that. A control unit that stores in a storage means the difference information between the image taken by the camera unit and the state and the standard state of the paper piece passing through the transport mechanism unit together with the handling data when the paper piece is handled.
A piece of paper handling device, characterized in that it is provided with.

[Appendix 2]
The transport mechanism unit includes a movable portion that is in charge of transporting a piece of paper and a device control unit that controls the movable portion.
The learning unit derives a compensating value that brings the transport closer to the standard state according to the state of the piece of paper input as learned data by machine learning.
The device for handling a piece of paper according to Appendix 1, wherein the device control unit controls the movable part based on the compensating value for bringing the loaded piece of paper closer to the standard state based on the machine learning result.

[Appendix 3]
The control unit extracts the difference between the state of the piece of paper passing through the transport path and the standard state by pattern recognition for each factor.
The learning unit learns the compensation value that minimizes the difference from the standard state for the image pattern related to the transportation of the paper piece for which no abnormality is detected together with the handling data as learned data.
The transport mechanism unit can control each of the feeding gap, the medium pressing force, the transport speed, the guard time, and the movable mechanism of the accumulation area, and each movable portion is set to the state of transporting a piece of paper in real time according to the learned data. The paper piece handling device according to Appendix 2, which is also optimized by the control unit.

[Appendix 4]
The transport mechanism unit includes a loading port for loading paper pieces by the user and a paper piece collecting unit for accumulating the loaded paper pieces.
The camera unit photographs the loading unit, the transport path, and the paper piece accumulating unit each time a piece of paper is handled.
The learning unit includes a state of a piece of paper loaded into the loading port, a state of a piece of paper passing through the transport path, and a state of a piece of paper accumulated in the paper piece accumulating unit, respectively, taken by the camera unit. The paper piece handling device according to any one of Supplementary note 1 to 3, wherein the difference information from the standard state of the above is machine-learned together with the handling data.

[Appendix 5]
The learning unit is extracted as a difference from the standard state, that is, the outer shape, the bending habit, the length in the thickness direction, the set deviation, the foreign matter mixed, the foreign matter adhered, the sluggishness, the distance between the media, the skew, the accumulated state, and the accumulation. The paper piece handling device according to any one of Supplementary note 1 to 4, wherein the empty area, the generated fold, the feeding time, the medium skew ratio, and the deformed shape are learned as factors.

[Appendix 6]
The piece of paper is a banknote and
The camera unit captures a state in which the user sets the bills in the slot, a state in which the bills are deteriorated, a state in which the bills are delivered from the bill slot, a state in which the bills are transported, and a state in which the bills are accumulated, each time the bills are handled. ,
In the machine learning, the setting state by the user in the bill insertion slot and the deterioration status of the bill are used as customer-side factor information, and the device side sets the bill feeding state, the bill transporting state, and the bill accumulation state as the customer side factor information. The paper piece handling device according to any one of Appendix 1 to 5, wherein the factor information is repeatedly machine-learned.

[Appendix 7]
The learning unit extracts the pattern of the state that causes an error and the state before the error by leveling as a factor related to the transportation of a piece of paper regarding the difference from the standard state for machine learning.
4. Paper piece handling device.

[Appendix 8]
Item 1. The described paper piece handling device.

[Appendix 9]
At least a picture of the transport path of the piece of paper inserted by the user in the device and the loading section where the user inserts the piece of paper. The state of the piece of paper that causes an error from the handling data is machine-learned by the learning unit in advance as learned data, and then
When handling a piece of paper, a process of detecting feature matching by comparing the learned data generated by machine learning with the state of the piece of paper that has passed through the transport path this time is executed.
When an error occurs when handling a piece of paper, or when the learned data generated by the learning unit is compared with the state of the piece of paper that has passed through the transport mechanism unit this time and the characteristics match, before and after that. The storage means stores the difference information between the image taken by the camera unit and the state and the standard state of the piece of paper passing through the device together with the handling data when the piece of paper is handled.
A method of controlling a piece of paper handling device, which is characterized in that.

[Appendix 10]
A transport mechanism unit that has a paper piece loading section and a transport path to transport the paper pieces inserted by the user, and a camera unit that at least captures the transport path in the apparatus and the loading section in which the user inserts the paper pieces. A computer for controlling a piece of paper handling device, which is equipped with
A learning unit that machine-learns the state of a piece of paper that causes an error from the state of the piece of paper passing through the transport mechanism unit and the handling data collected by taking a picture each time the piece of paper is handled by the camera unit as learned data.
When an error occurs when handling a piece of paper, or when the learned data generated by the learning unit is compared with the state of the piece of paper that has passed through the transport mechanism unit this time and the characteristics match, before and after that. A control unit that stores in a storage means the difference information between the image taken by the camera unit and the state and the standard state of the paper piece passing through the transport mechanism unit together with the handling data when the paper piece is handled.
A program characterized by operating as.

1 Paper piece handling device 2 Cash processing device 10 Transport mechanism unit 11 Moving unit 12 Equipment control unit 20 (20-1 to 20-n) Camera unit 30 Learning unit 40 Control unit 41 Processor 42 Image data acquisition unit 43 Pattern recognition unit 100 ATM
101-103 Other ATMs
110 Data server 120 Management department 121 Maintenance department 122 Technical department 123 Data analysis department

Claims (10)

A transport mechanism unit that includes a paper piece loading port and a transport path to transport the paper pieces loaded into the loading port by the user, and a transport mechanism unit.
A first camera unit that captures at least an image of the slot and acquires an image showing the state of the piece of paper caused by the user loaded from the slot.
Each time the piece of paper is handled, the user-side factor information caused by the user input from the input port is extracted from the image group captured by the first camera unit, and the user-side factor information is extracted from the user-side factor information. A learning unit that machine-learns the state of a piece of paper as learned data,
When the learned data generated by the learning unit is compared with the user-side factor information at the time of user input and the characteristics match, the difference between the state of the paper piece inserted from the input port and the standard state. A control unit that associates information with the corresponding image and stores it in the storage means,
A piece of paper handling device, characterized in that it is provided with. The paper piece handling device according to claim 1, wherein the user-side factor information includes information representing a set state of the paper piece and a deteriorated state of the paper piece with respect to the transport mechanism unit. It has a second camera unit that photographs the transport path and acquires an image showing the transport state of the paper piece caused by the device.
Each time the learning unit handles the piece of paper, the learning unit extracts device-side factor information due to the device state from the image group captured by the second camera unit, and from the user-side factor information and the device-side factor information. The paper piece handling device according to claim 1 or 2, wherein the state of the paper piece causing an error is machine-learned as learned data. The learning unit machine-learns images of the state of the paper pieces passing through the input port and the transport path taken by the first camera unit and the second camera unit, which are collected every time the paper piece is handled, together with the handling data. The paper piece handling device according to claim 3, wherein the paper piece handling device is characterized by the above. The transport mechanism unit includes a plurality of movable parts that are elements responsible for transporting the paper piece and includes a movable mechanism, and a device control unit that controls each movable mechanism included in the movable part.
The control unit extracts the difference between the state of the piece of paper passing through the transport path and the standard state by pattern recognition for each factor.
The learning unit learns the compensation value that minimizes the difference from the standard state for the image pattern related to the transportation of the paper piece for which no abnormality is detected together with the handling data as learned data.
The device control unit can control the movable mechanism of the movable unit, respectively.
3. The control unit is characterized in that, with reference to the compensation value, the movable mechanism of each of the movable parts is optimized in real time according to the state of transportation of the paper piece via the device control unit. Or the paper piece handling device according to 4. The transport mechanism unit includes a loading port for inserting the paper pieces by the user and a paper piece collecting unit for accumulating the loaded paper pieces.
A third camera unit for photographing the paper piece accumulating unit each time the paper piece is handled is provided.
The learning unit collects the state of the paper piece inserted into the loading port, the state of the paper piece passing through the transport path, and the state of the paper piece taken by the first to third camera units, and the paper piece accumulating unit. Machine learning of the state of the piece of paper and the difference information from each standard state together with the handling data is performed.
The learning unit is extracted as a difference from the standard state, that is, the outer shape, the bending habit, the length in the thickness direction, the set deviation, the foreign matter mixed, the foreign matter adhered, the sluggishness, the distance between the media, the skew, the accumulated state, and the accumulation. The paper piece handling device according to claim 4, wherein any or all of the empty area, the generated fold, the feeding time, the medium skew rate, and the deformed shape are learned as factors. The piece of paper is a banknote and
Wherein in the first to third camera unit, the set state of the bill by the user into the insertion slot, the bill deteriorated state, the bill feed-out state from the inlet, transfer state of the bill, the bill The accumulated state of the banknotes is photographed each time the banknotes are handled.
In the machine learning, the state in which the banknotes are set by the user in the slot and the state in which the banknotes are deteriorated are used as factor information on the user side, and the state in which the banknotes are delivered from the slot, the state in which the banknotes are conveyed, and the state described above. The paper piece handling device according to claim 6 , wherein the accumulated state of banknotes is used as factor information on the device side for repeated machine learning. Any one of claims 1 to 7, wherein at least one of each photographing point used for machine learning in the transport mechanism unit is increased by using a reflecting surface having a mirror function and / or a fisheye lens. The paper piece handling device described in 1. The user-side factor information caused by the user is extracted from the image group collected by taking a picture each time the paper piece is handled by the first camera unit that at least takes a picture of the slot for inserting the piece of paper, and the use thereof. The state of the piece of paper is machine-learned in advance as learned data from the factor information on the person side, and then
When handling the piece of paper, a process of comparing the learned data generated by the machine learning with the state of the piece of paper that has passed through the transport path this time to detect feature matching is executed.
When the learned data is compared with the user-side factor information at the time of user input and the characteristics match, the difference information between the state of the paper piece inserted into the input port and the standard state is converted into the corresponding image. A control method for a piece of paper handling device, characterized in that it is associated and stored in a storage means. The transport mechanism unit that includes a paper piece loading port and a transport path for transporting a piece of paper that the user has loaded into the loading port, and at least the loading port that has been photographed and is caused by the user that has been loaded into the loading port. A computer for controlling a piece of paper handling device provided with a first camera unit for acquiring an image showing the state of a piece of paper.
Each time the piece of paper is handled, the user-side factor information caused by the user input to the input port is extracted from the image group captured by the first camera unit, and the user-side factor information is used to obtain the above-mentioned information. A learning unit that machine-learns the state of a piece of paper as learned data,
When the learned data generated by the learning unit is compared with the user-side factor information at the time of inputting the user and the characteristics match, the state of the piece of paper inserted into the input port and the standard state A control unit that associates difference information with the corresponding image and stores it in the storage means,
A program characterized by operating as.

Images